INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

aughs

-0.75

 Spokane

-0.71

 Proposition

-0.67

cffff

-0.66

udeau

-0.66

yip

-0.65

#$

-0.64

 Doctrine

-0.63

 Contrast

-0.62

 Citiz

-0.61

POSITIVE LOGITS

ente

0.89

cipline

0.85

male

0.82

 princip

0.74

enf

0.70

 Alto

0.70

setting

0.68

rave

0.66

toc

0.66

Man

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.